[geeks] nerd reading for a Friday night ... old-skool waxed

[Charles Shannon Hendrix]

Wed, 31 Jan 2007 @ 18:10 -0500, der Mouse said:

> > Did uucp maps have a mechanism for expressing how quick and reliable
> > particular links were ?
> 
> Yes, though not a particularly good one.  (There was simply a
> preference value, which gave no indication of the speed/reliability
> tradeoff - a fast but flaky link might well have the same preference
> value as a slow but solid link.)

Taylor did more than that.

> > I would guess that without that, email over uucp could be very
> > variable in reliability and speed.
> 
> Even with that, it could. :-)

Even today, it does.

In fact, when I moved from UUCP to SMTP for email, my email reliability
went down and my connect time went up.

UUCP was efficient because it operated on compressed batches.

Almost all other protocols are very inefficient.

I don't see why they don't implement compressed batch SMTP.

The advantages of UUCP and SMTP all in one.

> Yes, in the early days of the mapping project.  Later, when the maps
> got posted to (IIRC) comp.uucp.maps, you could set it up to feed new
> maps into your local pathalias database automatically.

People also distributed maps with rdist, email, UUCP, and all kinds of
ways.

It wasn't a perfect system, but I can't help but notice that a lot of
VPNs remind me of the old days of thriving private UUCP networks.

-- 
shannon "AT" widomaker.com -- ["If you tell the truth, you don't have to
remember anything" -- Mark Twain]